Retrospective analysis of colistin-resistant bacteria in a tertiary care centre in India

The first report of a multiclonal cluster of Klebsiella pneumoniae was published by Antoniadou et al1 in 2007 from a Greek intensive care unit (ICU). They reported 18 isolates of K. pneumoniae in 13 patients. There were six distinct clones identified by repetitive extragenic palindromic polymerase chain reaction (REP-PCR). A cluster of five cases of carbapenem-resistant colistin-resistant K. pneumoniae was identified at Detroit Medical Center2. Genotyping in this cluster revealed two closely related clones which indicated a strong link for patient-to-patient transmission2. Ghafur et al3 reported such pan drug-resistant (PDR) organisms in 13 patients from India. Colistin is an old antibiotic which has resurfaced in past few years due to rising incidence of multidrug resistance organisms3. There have been reports of colistin-resistant and tigecycline-resistant Acinetobacter baumanii (resistant to both tigecycline and colistin) from north India4 which was 3.5 per cent of total and 16 per cent of all MDR organisms isolated from urine samples. Prior colistin exposure has been considered as an important association though not established in literature345678910.

We did a retrospective analysis of the case records of the patients who had any culture report positive for PDR organisms between May 1, 2016 and January 31, 2017 admitted in the three ICUs of Apollo Hospitals, Bhubaneswar, India. An organism was considered PDR if it was resistant to all antipseudomonal agents such as penicillins, cephalosporins, carbapenems, aminoglycosides, quinolones, monobactams and polymyxins. Cultures were done using VITEK-2 compact (Biomerieux India Pvt. Ltd, software version 8.01) for all specimens. The report of VITEK-2 and European Committee for Antimicrobial Susceptibility Testing (EUCAST) breakpoints (S≤2, R≥2)³ were followed for Enterobacteriaceae, Pseudomonas and Acinetobacter isolates which were considered resistant to colistin if the minimum inhibitory concentration (MIC) was ≥8 and 4 μg/ml, respectively, as per the Clinical and Laboratory Standards Institute (CLSI) guidelines3.

The reports of colistin resistance were confirmed by e-strip method [Colistin Ezy MIC^™ Strip (0.016-256 mcg/ml) from HiMedia Laboratories Pvt. Ltd., Mumbai] in these isolates. Determination of an isolate as an infective pathogen or colonizer was done by two independent investigators who were blinded to the study design. The Acute Physiology Assessment and Chronic Health Evaluation (APACHE) II score on admission for all patients, their culture reports, antibiotics received and length of stay (LOS) in hospital along with outcome were analyzed.

There were a total of 1448 ICU admissions in three ICUs of the hospital during the study period. There were a total of 15 isolates of PDR organisms from 11 patients. There were six male and five female patients. Detailed information of these patients is summarized in the Table. Average length of hospital stay was 60.72 (range 25-123) days. The mean age of the patients was 61.8 (28-84) yr. K. pneumoniae was isolated from 10 cases, Pseudomonas from four (P. aeruginosa in 3 and P. luteola in 1 case) and A. baumanii from one case. These isolates were considered to be colonizer in five and infective pathogens in six patients. These PDR isolates were associated with catheter-associated urinary tract infection (7), tracheitis (4), bacteraemia (2), meningitis (1) and soft-tissue infection (1). Average APACHE II score was 24 (7-36) indicating sick patients with multiple co-morbidities and organ dysfunction. Nine patients had received prior (i.e., before the development of PDR infections) colistin therapy due to the growth of multidrug-resistant bacteria isolates while two had no history of prior colistin treatment. Of the nine patients, eight had combined colistin therapy while one had colistin monotherapy.

Nine of these patients died while the two were discharged. The overall mortality (9/11, 81.81%) was due to the primary disease process and other co-morbid conditions. The high APACHE II scores of all the patients who died indicated the severity of the illness and co-morbid conditions. Two of the nine deaths were attributed to PDR infections (infection-related mortality 18.18%). One patient with traumatic brain injury (TBI) died due to pulmonary embolism after discharge from the hospital. Eight of the 11 patients had neurological issues (stroke and TBI) and had prolonged hospitalization. In two reports from a Greek ICU by Falagas et al56, the overall death rate was 29 and 41.7 per cent, respectively. The infection-related mortality in the second series was 33.3 per cent6. In the study reported by Tsioutis et al7, mortality was 23.8 per cent (5 of 21). In these studies, PDR infections were treated with colistin along with carbapenems and quinolones567. Duration of treatment was between 8 and 34 days in these studies56. Except one, all our PDR-infected patients were treated with polymyxin along with third-generation cephalosporins or carbapenems for 14-20 days. Polymyxin was preferred over colistin due to its presumed lower nephrotoxicity11 as most patients had deranged renal parameters. In the previous studies, the patients were treated with colistin even though the isolates had in vitro resistance to the same drug567. Tsioutis et al7 implicated a better cure rate with tigecycline in PDR cases than colistin. None of our patients were given tigecycline.

Multidrug-resistant organisms were present in nine of 11 patients before the development of PDR organisms. Three out of four colonizers were treated with colistin. The duration of colistin therapy was between 10 and 14 days. The growth of PDR organisms in two patients, who did not have any prior exposure to colistin, was thought to be transferred from other infected patients. The genetic analysis could not be done which was a limitation of our study. Increased exposure to colistin has been viewed as the single most important risk factor for the development of colistin resistance among Gram-negative bacteria in the previous studies1891012.

The average length of stay (LOS) was around 60 days which matched with other studies8101112. In such critical cases, the average LOS before PDR culture report was 34.18 days (11-70 days). The time gap between the identification of PDR infection and death in the nine patients varied between 8 and 53 days. All patients with PDR infections had long LOS with high APACHE II scores.

It is important that culture reports should be judiciously interpreted to differentiate between colonizer and infective pathogen before the treatment. Widespread indiscriminate use of colistin to treat carbapenem resistant Enterobacteriacae (CRE) and other Gram-negative organisms can lead to the emergence of PDR organisms. Strict implementation of antibiotic stewardship programme is essential to limit use and prevent abuse of colistin. A large prospective study is required to delineate the risk factors for such PDR infections in Indian hospitals.